DOI: 10.5281/zenodo.

10603960

IMPACT OF SOCIAL CAPITAL ON UNIVERSITY STUDENTS'

INNOVATIVE PERFORMANCE: EXPLORING CREATIVE
COGNITIVE STYLE AS A MEDIATING VARIABLE

QINGHONG LIN
Doctoral Student, College of Innovation and Management, Suan Sunandha Rajabhat University, Bangkok,
Thailand. Email: 371751249@qq.com
NIYOM SUWANDEJ
College of Innovation and Management, Suan Sunandha Rajabhat University, Bangkok, Thailand.
Email: niyom.su@ssru.ac.th

Abstract
The quest for innovation plays a pivotal role in advancing scientific and societal progress. Higher education
endeavors to nurture students' innovation capacities, fostering critical thinking, astute judgment, and the courage
to challenge established norms while embracing human civilization's achievements. This study delves into the
influence of social capital within social networks on college students' innovative performance. This paper aims to
investigate the impact of social capital on college students' innovation performance, discerning its dimensions and
mediating effects. It constructs a structural equation model to explore the relationship between college students'
social capital (network ties, trust, shared goals) and their innovation performance dimensions (motivation,
thinking, personality, achievement). The empirical validation involves a sample of 480 participants. The study
delineates college students' social capital and innovation performance into specific dimensions and employs a
structural equation model to gauge their interrelation. Social capital comprises network ties, trust, and shared goals,
while innovation performance is assessed through motivation to innovate, thinking ability, personality traits, and
actual innovation achievements. Empirical validation through structural equation modeling confirms a positive
correlation between college students' social capital and innovation performance. Moreover, it highlights the
mediating role of college students' cognitive innovation style in this relationship. This scholarly inquiry not only
deepens our comprehension of the multifaceted factors influencing college students' innovation but also provides
vital insights for educational policy formulation, pedagogical enhancements, and the cultivation of students'
innovative prowess.
Keywords: Social Capital, Innovation Performance, Teacher Innovation Support.

1. INTRODUCTION
Innovation is the cornerstone of scientific progress and social advancement, and therefore the
cultivation of the innovative capacity of university students is a fundamental goal of higher
education. Fostering students' capacity for innovation involves developing their ability to think
rationally, make sound judgments and analyze critically, and encouraging them to challenge
existing paradigms on the basis of human achievements. This entails questioning established
knowledge, embracing originality, exploring and innovating scientific fields, and thus
contributing to the continuous progress of society.
China's keen focus on "double creativity" and its promotion of innovation and entrepreneurship
has earned unprecedented attention. Scholarly discourses have emphasized that innovation is

1156 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

an important driver of productivity (Cainelli et al., 2006; Love & Roper, 2015), sustained
economic growth (Zhi & Shudan, 2015), and firm performance (Hou et al., 2019). Notably,
Baer et al. (2003) emphasized the important impact of creativity on employee innovation
performance. However, the evolving innovation landscape has become intricate, resource-
intensive, and risky (Dziallas & Blind, 2019), which has implications for social capital on
college campuses.
Understanding the factors and mechanisms that influence innovation among university students
is critical to the development of innovators. Exploring these aspects can help educational
authorities to formulate relevant policies as well as universities to reform their educational
frameworks. This study provides insights into the impact of social capital on university
students' innovative performance and also explores the mediating role of innovative cognitive
styles.
This study specifically focuses on the innovative performance of university students in
university research or practical innovation activities, as distinct from the work-related
innovation of corporate employees. It aims to address challenges, accumulate knowledge,
develop problem-solving skills, and generate tangible and valuable innovations through novel
insights.
The goal of the study is to investigate the direct effect of social capital on college students'
innovative performance and the mediating role of innovative cognitive style in the effect of
social capital on innovative performance.

2. LITERATURE REVIEW AND HYPOTHESIS

2.1 Definitions of Key Terms
Social capital: Nahapiet and Ghoshal's (1998) three-factor structural theory defines social
capital within various social contexts, comprising the structural, relational, and cognitive
dimensions. The structural aspect focuses on the breadth and strength of social networks.
Relational dimension emphasizes interpersonal connections formed over time, with trust being
fundamental in these implicit relationships, stemming from emotional and rational interactions.
The cognitive dimension acts as a resource enabling mutual understanding and shared
expression among network actors, employing common goals and language for definition.
Creative cognitive style: Individual differences in organizing and processing information
consistently, influencing problem comprehension and resolution. This study employs the
Adaptive Innovation Scale (KAI inventory) developed by M. Kirton (1976) to measure
cognitive styles.
Innovative performance: Efforts to meet the challenges of scientific research or innovation by
accumulating knowledge, gaining insights, developing problem-solving skills, and producing
valuable results. It consists of two components (Amabile, 1997): intrinsic capabilities and
innovative outcomes. Intrinsic competence is related to the physical, intellectual and thinking
factors that stimulate innovation, mainly in terms of students' motivation to innovate,

1157 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

personality and thinking. Innovative outcomes are recognized novel and practical results, as
reflected in publications, conference participation, patents and competition results (McWilliam,
2009).
The modeling assumptions are shown below.

Figure 1: Conceptual Framework: The influence of social capital on college students'

innovation performance
2.2 Conceptual model and Hypotheses
Hanna Rydehell (2019) explored the relationship between innovation performance and external
financing capabilities by examining patents and product differentiation. Wu Bohong (2018)
uses indicators to measure innovation inputs and outputs in Huai'an City and seeks ways to
improve innovation performance based on empirical data. Yang Xu (2019) investigated the
innovation performance problem of DT companies and proposed a solution by drawing on
information asymmetry and R&D management theory. Wang Yi (2020) studied the innovation
performance of Haier Group in six years through industry comparison and assessed its
innovation process and output.
College students' social capital has an important impact on innovation performance in two ways.
First, it promotes motivation to innovate and develops innovative personality traits and
innovative thinking, thus improving innovation outcomes and overall performance. A sound
social network of college students plays a pivotal role in cultivating innovation ability,
facilitating the exchange of innovative ideas, and significantly enhancing innovation
motivation, innovative personality, innovative thinking ability, and innovation outcomes.

1158 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

3. RESEARCH METHODOLOGY
This part of the researchers used a set of questionnaires to validate the model and hypotheses,
aiming to explore the relationship between social capital of university students innovation
performance of university students, as well as their internal mechanisms and boundary
conditions. The finalized model will be used for data analysis and interpretation.
3.1 Pre-survey
In the pre-survey stage, 58 questionnaires were distributed. After excluding invalid
questionnaires, the valid questionnaires of the pre-survey were 51, and the effective recovery
rate of the samples was 87.93%. After the reliability analysis of the pre-survey, "Cronbach
Alpha if Item Deleted", CITC (Corrected Item-Total Correlation), item analysis, and expression
accuracy or not of the four analytical treatments, and after the expert's advice, retained 44 items,
the model relationship and items are shown below figure 2.
3.2 Population and Sample
The data came from universities in the top three largest cities in Guangxi Province China,
covering different majors and different grades. The questionnaires were collected online, and
finally 540 responses were obtained, and after deleting invalid and incomplete questionnaires,
the number of valid questionnaires was 480, with a validity rate of 88.88%.

Figure 2: Model diagrams and their question items identified after the pre-survey

1159 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

3.3 Instrument
Of all the scales, the study used a 5-point Likert subscale, with 5 indicating agreement and 1
indicating complete disagreement.
Table 1: Specific references to the scale sources and included question items
variable& Encoding Measurement Dimension Encoding of items References
Innovation Motivation IM1-3
Amabile（1997）
Innovation Innovation Personality IP1-3
(Liu & Fan, 2020)
Performance (IPs) Innovation Thinking IT1-4
(Kukkonen & Bolden, 2022)
Innovation Achievement IA1-3
Network Tie SNT1-4 (Nahapiet & Ghoshal, 1998)
Trust STR1-5 Chow&Chan（2008）
Social Capital (SCs)
(Zhang Y., 2014)
Shared Goal SSG1-4
(Wang et al., 2020)
Originality CO1-6 Kirton(1976)
Creative Cognitive
Efficiency CE1-6 (Zhang et al., 2018)
Style (CCs)
Rule CR1-6
Data source: Author's compilation based on references

4. DATA ANALYSIS AND RESULTS

The study employed quantitative analysis to present its findings. As the data originated from a
single source, the potential for common method variance (CMV) was acknowledged. To assess
CMV in the partial least squares structural equation modeling, two techniques were employed:
a comprehensive examination of covariance utilizing the variance inflation factor (VIF) as
proposed by Kock (2015) and a correlation matrix procedure. Neither of these methods
revealed any bias in the single-source data.
4.1 reliability and validity
It's important to highlight that Error! Reference source not found. presents both the first-
order and second-order variables simultaneously. Whether they are first-order or second-order,
the table indicates that these reflective dimensions exhibited Cronbach's α and CR values
exceeding 0.70. Additionally, all scales showcased AVE values surpassing 0.50, thereby
confirming the convergence validity of these dimensions/constructs.

1160 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

Table 2: Data Reliability and Convergent Validity Tables

Constructs and Items Cronbach's alpha CR Value AVE Value
Step I: First-order reflective components were evaluated.
IM 0.851 0.909 0.77
IP 0.847 0.907 0.766
IT 0.895 0.927 0.76
IA 0.861 0.915 0.783
SNT 0.894 0.926 0.759
STR 0.897 0.924 0.709
SSG 0.883 0.92 0.741
CO 0.924 0.936 0.725
CE 0.917 0.936 0.708
CR 0.899 0.922 0.664
Step II: The second-order reflective construct is presented here
IPs 0.771 0.853 0.593
SCs 0.751 0.858 0.668
CCs 0.721 0.843 0.642

For evaluating discriminant validity, HTMT criteria was utilized for the inner construct. All
HTMT values in the table are less than 0.85, implying that there is a good differentiation
between the three measures of social capital, innovation cognitive style and innovation
performance, and that the data under study has a good differentiation validity between the three
scales.
Table 3: HTMT Validity Test
CCs IPs SC
CCs -
IPs 0.61
SC 0.648 0.605 -

4.2 Hypothesis testing

The R-squared values of CCs, IPs and SFs are all between 0.33 and 0.67, indicating a moderate
interpretation.
The effect sizes compiled in this study are shown in the table below. Effect size (f2) , the effect
value influences the effect at a level of 0.35 (high), 0.15 (medium), and 0.02 (low) from high
to low (Cohen, 1988). The effect size of SCs>CCs is high (0.305), while the effect size of two
relationships are mid-to-low, which is not high but still within the critical values.
Table 4: Effect size (f2)
CCs IPs
CCs 0.106(mid-to-low)
SC 0.305(medium) 0.103(mid-to-low)
Q-square. Following the procedure suggested by Shmueli et al. (2016), the current
implementation of the PLSpredict algorithm in the SmartPLS software allows researchers to
obtain k-fold cross-validated prediction errors and summary statistics of prediction errors,

1161 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

which can be used to compare their prediction performance with two naïve benchmarks
(Shmueli et al., 2019).
(1) Q²value in PLSpredict compares the prediction error of a PLS path model to the simple
average prediction. If the Q² value is positive, the PLS-SEM model has better prediction
performance.
(2) The prediction error (e.g., RMSE or MAE) of the PLS-SEM results should be lower than
the prediction error of the LM results when compared to the LM results
(3) The prediction error of PLS-SEM results should be lower than that of LM results.
Table 5 provide the Q²predict values for the model's explicit variables (dimensions), and all
values are greater than 0, so the dimensions have predictive power. The PLS-SEM_RMSE <
LM_RMSE for most of the metrics indicates that the model has medium predictive power.
Table 4: PLSEpredict assessment of Manifest Variables
Q²predict PLS-SEM_RMSE PLS-SEM_MAE LM_RMSE LM_MAE
CE 0.136 0.931 0.796 0.935 0.801
CO 0.161 0.918 0.785 0.92 0.784
CR 0.15 0.923 0.782 0.926 0.783
IA 0.12 0.939 0.798 0.943 0.801
IM 0.142 0.928 0.781 0.931 0.783
IP 0.105 0.948 0.8 0.95 0.802
IT 0.134 0.932 0.789 0.935 0.789
In Table 6, the Q²predict values for the latent variables are all greater than 0, which indicates
that the PLS-SEM model is able to predict the latent variables with moderate to good accuracy.
The RMSE values and MAE values also indicate that the PLS-SEM model is able to predict
the latent variables with moderate to good accuracy.
Table 5: PLSEpredict assessment of Latent Variables
Q²predict RMSE MAE
CCs 0.229 0.88 0.722
IPs 0.211 0.891 0.739

4.2.1 Path analysis and hypothesis testing for structural modeling

The magnitude and significance of the path coefficients were used to assess the relationship
between the research constructs. The results of calculating the path coefficients and T-values
using Bootstrapping method are shown below Figure 3 and Table4.
In Figure 3, the results of the path analysis test demonstrate significant relationships in this
study. Both social capital (SCs -> IPs, β=0.314, p<0.001) and cognitive styles of innovation
(CCs -> IPs, β=0.311, p<0.001) exhibit a positive influence on innovation performance (IPs).
As a result, hypotheses H1 and H2 are supported and cannot be rejected. Moreover, hypothesis
H3, which posits a positive influence of innovation cognitive style (CCs -> IPs, β=0.311,
p<0.001) on innovation performance, is also supported. Thus far, all direct effects in the model
have been validated, confirming the support for hypotheses H1-H3.

1162 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

Figure 3: Path relationship

4.2.2 Mediation Testing Hypothesis
In this study, the mediation role of CCs in the relationship between the variables SCs with IPs
was examined using the Bootstrap mediation effect test. The test was conducted with a 95%
Bias Corrected confidence interval and 5000 repetitions to determine the significance of the
mediation effect. The results of the mediation effect test are presented in Table 7. total effect
for SCs -> IPs (β=0.464, p<0.001), The obtained p-values were below the critical values for
statistical significance, 0.001, with values surpassing the critical value of 1.96, indicating
statistical significance. These findings provide substantial evidence supporting the mediating
effect of CCs.
Table 7: Total effects and specific indirect effects- Mean, STDEV, T values, p values
Original Sample Standard deviation T statistics P
sample (O) mean (M) (STDEV) (|O/STDEV|) values
CCs -> IPs 0.311 0.31 0.047 6.571 0
SCs -> CCs 0.484 0.485 0.035 13.691 0
SCs -> IPs 0.464 0.466 0.037 12.635 0
SCs -> CCs -> IPs 0.15 0.15 0.025 6.021 0
The analysis from Table 7 highlights specific mediation pathways. SCs -> CCs -> IPs path,
totaling an indirect effect value is 0.464. The mediation effect of SCs -> SFs -> IPs is 0.15,
accounting for 32.32% of the total indirect effect. This confirms Hypothesis 4, indicating a
partial mediator effect due to the presence of a direct effect from SCs to IPs.

1163 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

5. CONCLUSION AND DISCUSSION

5.1 Conclusion
In this study, the proposed hypotheses were thoroughly tested using quantitative analysis
techniques. The results of the study indicate that there is a significant relationship between
college students' social capitals (SCs) and their innovation performance (IPs), as well as
between college students' cognitive styles of innovation (CCs) and their innovation
performance (IPs). In addition, college students' social capital has a significant effect on their
cognitive styles of innovation (CCs).
The mediating role of innovation cognitive styles (CCs) between social capitals (SCs) and
innovation performance (IPs) was also confirmed, affirming some of its mediating effects.
Moreover, social capital has a relatively strong level of predicting cognitive style of innovation
(KAI) and innovation performance. The results of the study emphasize the urgent need for
educational institutions and instructors to prioritize the development of students' social
networks and psychological resilience in order to promote the comprehensive and holistic
development of their innovative capabilities.
5.2 Discussion
The study's results underscore the influential role of both social capital and innovative cognitive
style in shaping university students' innovation performance. The positive impact of social
capital on fostering innovative cognitive styles among students highlights the significance of
robust social networks and exposure to diverse knowledge in enhancing creativity. Moreover,
the mediating effect of innovative cognitive style signifies its pivotal role in channeling the
influence of social capital onto innovation performance. These findings offer essential insights
into the complex interplay between social factors and individual cognitive attributes in driving
innovation among students.
5.3 Limitations and Future Perspectives
While the study offers valuable insights, there remains an avenue for refinement in terms of
sample diversity and data depth. Augmenting the breadth of the sample pool and diversifying
data collection methodologies could substantially enrich our understanding of the multifaceted
factors that wield influence over innovation capabilities. Moreover, the contextual specificity
of the study warrants its validation across diverse cultural and social milieus.
Prolonged longitudinal studies could unveil the enduring impacts of social capital and faculty
support on innovation prowess. Further investigations exploring additional mediating variables
could unravel nuanced layers of the impact of social capital and faculty guidance.
Consideration of comparative studies spanning students from myriad backgrounds might
illuminate a more comprehensive spectrum of influential factors at play in shaping innovation
capacities.

1164 | V 1 9 . I 0 1
 DOI: 10.5281/zenodo.10603960

References
1) Abou-Shouk, M., Zoair, N., Aburumman, A., & Abdel-Jalil, M. (2022). The effect of personality traits and
knowledge-sharing on employees’ innovative performance: A comparative study of Egypt and Jordan.
Tourism Management Perspectives, 44, 101024.
2) Amabile, T. M. (1997). Motivating creativity in organizations: On doing what you love and loving what you
do. California Management Review, 40(1), 39–58.
3) Baer, M., Oldham, G. R., & Cummings, A. (2003). Rewarding creativity: When does it really matter? The
Leadership Quarterly, 14(4–5), 569–586.
4) Cainelli, G., Evangelista, R., & Savona, M. (2006). Innovation and economic performance in services: A
firm-level analysis. Cambridge Journal of Economics, 30(3), 435–458.
5) Chow, W. S., & Chan, L. S. (2008). Social network, social trust and shared goals in organizational knowledge
sharing. Information & management, 45(7), 458-465.
6) Dziallas, M., & Blind, K. (2019). Innovation indicators throughout the innovation process: An extensive
literature analysis. Technovation, 80, 3–29.
7) Kirton, M. (1976). Adaptors and innovators: A description and measure. Journal of Applied Psychology,
61(5), 622-629.
8) Kirton, M. J. (1994). Adaption-innovation: In the context of diversity and change. Routledge
9) Kukkonen, T., & Bolden, B. (2022). Nurturing Creativity in the Visual Arts Classroom Understanding
Teacher Strategies through Amabile's Componential Theory. Canadian Review of Art Education, 49(1), 46-
62.
10) Kock, N. (2015). Common method bias in PLS-SEM: A full collinearity assessment approach. International
Journal of e-Collaboration (ijec), 11(4), 1-10.
11) Liu F., Hu C., & Yi Y. (2022). The Effects of Epistemological Beliefs and Teachers' Innovation Support on
College Students' Creativity. In Journal of Yuncheng College (CNKI; Vol. 40, Issue 03, pp. 78-86).
12) Nahapiet, J., & Ghoshal, S. (1998). Social Capital, Intellectual Capital, and the Organizational Advantage.
Academy of Management Review, 23, 242–266.
13) Rydehell, H., Isaksson, A., & Löfsten, H. (2019). Business networks and localization effects for new Swedish
technology-based firms’ innovation performance. The Journal of Technology Transfer, 44(5), 1547–1576.
14) Wang, Y., & Yuan, L. (2020). A study on the impact of user participation on corporate innovation
performance (CNKI) [Master's degree, 10.27332/d.cnki.gshzu.2020.000534]. Shihezi University.
15) Wu, B. (2018). Research on the impact of R&D investment on corporate performance. In China Business
Journal (CNKI; Issue 06, pp. 89-90).
16) Yang, X. (2019). Research on R&D investment intensity and innovation performance in DT companies
(CNKI) [Master, 10.27362/d.cnki.gtsxy.2019.000013]. Tianjin University of Commerce.
17) Zhang C., Yu W., & Yao Q. (2022). Influence of mentor autonomy support on engineering graduate students'
innovative behaviors: a survey based on 12 "double first-class" universities. In Degree and Graduate
Education (CNKI; Issue 10, pp. 42-48).
18) Zhang H., Wang L., & Zhang M. (2018). Relationship between creative cognitive style, creative personality
and creative thinking. Psychological and Behavioral Research, 16(1), 51.
19) Zhang, Y. (2014). Research on the impact of social capital on graduate students' innovation ability doctoral
dissertation, University of Science and Technology of China).
20) Zhi, S., & Shudan, X. (2015). An evaluation of Chinese technological progress and the convergence of
economic growth-from the perspective of innovation and efficiency [J]. Social Sciences in China, 7, 4–25.

1165 | V 1 9 . I 0 1